PhD Opportunities

Project title: Probabilistic machine learning analysis of electrochemical data for characterization of mixed-species biofilms 

Biofilms are microstructured microbial communities that thrive at surfaces and interfaces. Differently from the planktonic lifestyle, in which single cells swim independently in liquid media, unaware of the presence of surrounding surfaces, biofilm “mode of life” entails a broad range of interactions among cells and between cells and the environment. The combined effect of these dynamic interactions is the aggregation of cells at the interface and the production of extracellular polymeric substance (EPS) that

keep the cells close to each other and attached to surfaces. In real-world, biofilms are open systems, so they are naturally exposed to influx and contamination from the surrounding environment. The natural consequence is that most biofilms are mixed-species communities, comprising bacteria and fungi. Mixed-species biofilms are a serious concern in healthcare, as they result in difficult-to-treat infections and they harbour antimicrobial resistant microorganisms. 

While natural biofilms comprise of multiple bacterial and fungal species, most studies still concern single species biofilms, which present an unrealistic response to antimicrobial agents, thus requiring costly animal model experiments for further validation. Recently, there have been several attempts to study multispecies biofilms using stable multispecies models. While the results obtained from these experiments are closer to reality, there is still a knowledge gap in the understanding of mixed-species

biofilm in open systems. Even if high-end methods like sequencing are available to characterize mixed-species biofilms, these are expensive and require highly skilled operators. Currently, there is not yet a single technology to monitor mixed-species biofilms, and complex, customized protocols are required to understand these complex systems in laboratory and industrial research. 

The availability of low-cost methods to characterize mixed-species biofilms will allow the rapid identification of pathogens in mixed-species biofilm infections and in water systems, with great benefits in terms of health and lower maintenance expenses, respectively. Further, the rapid identification of pathogens in the periprosthetic liquid or on the surface of the implant will enable effective prevention and treatment of serious infections, like those occurring in prosthetic and implanted patients. The global economic burden of biofilms is estimated in excess of $5000 bn a year. Overall, there is an urgent need for rapid and low-cost methods for mixed-species biofilm characterization in biomedical, environmental and bioprocess industry. 

Direct electrochemistry of biofilms is an established research area for biofilms characterization. Biofilm electrochemistry can contribute to the resolution of mixed-species biofilms, due to its low cost, real-time and non-destructive characteristic. While biofilm electrochemistry cannot provide a final identification of each microbial species, it is in theory possible to analyse the specific signature of

each microbial species using probabilistic machine-learning (PML) methods. In this project, the two PhD students will develop a novel method for real-time, online characterization of mixed-species biofilms using bioelectrochemical methods in combination with PML driven data analysis. The proposed method is expected to produce a digital output that can be directly analyzed and further elaborated at low cost. This is a step toward a more complete characterization and knowledge of mixed-species biofilms, which will have strong applications in health and industrial sectors. 

PhD student 1 will focus on the electrochemical, microscopy and spectroscopy characterization of polymicrobial biofilms. S/he should have a background in Chemical/Biochemical Engineering or Physical Chemistry, with a strong interest in Microbiology. This project will involve the collaboration of two external scientists, Massimiliano Galluzzi (SIAT-Shenzhen, China) and Elia Marin (KIT, Kyoto, Japan), which will contribute to the Atomic Force Microscopy (AFM) and Raman Spectroscopy characterization of the mixed-species biofilms, respectively. The PhD student 1 might be shortly seconded to these labs for training and part of the experimental work. 

PhD student 2 will focus on the writing of the Probabilistic Machine Learning (PML) code and electrochemical data (from PhD student 1) analysis for the modelisation of the mixed-specie biofilms. S/he would have a background in Data Science/Computer Science/Physics/Applied Mathematics/Engineering with a strong interest in Mathematical and Computational Biology. The PhD student 2 will work closely with the PhD student 1 to optimize the data acquisition pipeline. The PhD student 2 will be co-supervised by Alberto D’Onofrio, a PML and Mathematical/Computational Biology expert at University of Trieste, Italy. The project might include a short-term secondment at University of Trieste for training in mathematical methods and data analysis. 

Contact points

Enrico Marsili (UNNC) – [email protected]  

Elia Marin (Kyoto Institute of Technology, Japan) - [email protected]

Massimiliano Galluzzi (Shenzhen Institute of Advanced Technology, China) - [email protected]

Alberto D’Onofrio (University of Trieste, Italy) - alberto.d'[email protected]